Gfftk

* Upgrade the consensus module, more accurate calls and opposite strand filtering.
* A few bug fixes.

Bug fix release:

- fix fasta parser for whitespace 2
- fix genbank cleanup

- move to CalVar version system (date)
- update GFF3 parsers
- provide support for genbank flat file output

- fix for writing gzip output files

- add support for io.BytesIO objects to be used as input -- this won't have an effect on the cli `gfftk` package, but could benefit as a library if streams are used.

- Add support for detecting pseudogenes, ie mRNA features that have CDS with internal stops or those that are previously labeled pseudogenes. Pseudogenes will still be written, however, CDS will not be.
- Add option to `gff.gff2dict` for `gff_format="auto"` where the format will try autodetect to parse some different GFF formats. Currently just supports the default and the NCBI-eukaryote format.